[disgruntledphd2]

I mean, to be fair to OpenAI and this approach, new general purpose technologies (like LLMs) are normally pretty hard to work with at first.

Look at electricity in factories and how long it took to have an impact, as well as the massive, massive lag between Visicalc and economy wide productivity increases.

I personally do believe that the foundation model companies are a terrible business, but that the technology itself is definitely useful (even if it's vastly unlikely to go full Singularity).